Weekly Google Maps review triage to Slack and Linear

Build an agent workflow that monitors Google Maps reviews for our business locations every Monday morning, then produces two outputs: a Slack digest summarizing the week, and Linear tickets for recurring complaint themes the ops team needs to fix.

Trigger: cron, every Monday at 8am local time.

Step 1: scrape reviews. Use the Apify integration's 'Run Actor Synchronously and Get Dataset Items' operation to run a Google Maps reviews scraper Actor (default to apify/google-maps-reviews-scraper, but let the user swap it for any equivalent reviews Actor). Pass the configured list of Google Maps location URLs or place IDs and a date range covering the past 7 days. Persist the last successful run timestamp and the set of review IDs already processed in workflow memory, then filter the returned items so we only act on reviews newer than the last run, even if the Actor returns older ones.

Step 2: classify and group. For each new review, classify it as positive, neutral, or negative using both the star rating and the review text. For negative reviews, identify the dominant theme (food quality, wait times, cleanliness, staff behavior, value, service, accuracy, etc.). Do not rely on the scraper's built-in sentiment score, use the agent's own judgment so themes are concrete and actionable. Count theme occurrences across the week's batch and note which locations each theme shows up at.

Step 3: post the Slack digest. Use Slack Bot's 'Send a Message' operation to post to the configured channel. Include: overall sentiment breakdown for the week with counts and percentages; average star rating with a delta versus the prior week if memory has it; the top three complaint themes with mention counts; one or two representative review quotes per top theme; and a short callout for any 1-star review or unusual spike in negative volume. Use Slack mrkdwn formatting.

Step 4: file Linear tickets. For any negative theme that appears more than twice in this week's batch and does not already have an open ticket from a prior run, use Linear's 'Create Issue' operation to open a ticket in the operations team's backlog. Title is the theme phrased as a fix (for example, 'Wait times at the downtown location'). Description should include the mention count, the locations affected, and the supporting review quotes with reviewer names, star ratings, and review URLs. Default priority Medium. Track the set of open theme tickets in workflow memory so recurring themes get a comment with new quotes added to the existing ticket instead of a duplicate.

State to persist between runs: the last successful run timestamp, the set of review IDs already processed, and a map of negative themes to their open Linear ticket IDs.

If the Apify run fails or returns zero new reviews, still post a short Slack message noting the run status so the ops team knows the workflow ran. Make the location list, the Slack channel, the Linear team, the theme dictionary, and the ticket-creation threshold all configurable by the user when they install this.